Podcasts · The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence)

Relational Foundation Models for Enterprise Data with Jure Leskovec - #768

21 May 2026 · 1 h 06 min · Episode 768

In this episode, Jure Leskovec, co-founder and chief scientist at Kumo and professor of computer science at Stanford, joins us to explore two fronts of his work: AI for science and relational deep learning. We begin with AI Virtual Cell, a multiscale effort to learn data-driven representations from proteins to cells to patients using single-cell RNA-seq data, protein language models like ESM, and structure models like AlphaFold—without hand-encoding biology. Jure then dives into relational deep learning, reframing enterprise databases as graphs and training neural networks directly on raw multi-table data. He explains Kumo’s Relational Foundation Model (RFM2), which performs in-context learning over subgraphs to make accurate predictions on new databases and tasks with no training, and how this approach benchmarks against RelBench and other multi-table datasets. We also discuss real-world deployments at companies like Reddit, DoorDash, and Coinbase, explainability via attention over tables and columns, integration with agentic systems, deployment options, and practical limitations.

The complete show notes for this episode can be found at https://twimlai.com/go/768.

Links in the notes 1

• https://twimlai.com/go/768

The episode on the publisher’s site ↗ (opens in a new tab)